Thank you both. Simple, as expected. And I’m easily surprised.
Version 10 (perhaps obviously) for those scoring at home.
> On May 1, 2018, at 10:11 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>
> Rob Sargent <robjsargent@gmail.com> writes:
>> Should I be? I would have thought the pk would have been chosen v.
>> function index?
>
> If I'm reading this correctly, the PK index contains uuids while
> the fpv index contains float4s, meaning the latter is probably half
> the size. So scanning it is a lot cheaper, at least according to
> the planner's cost model.
>
> regards, tom lane